April 24, 2006 – The random noise in digital images may not be so random after all. Researchers from Binghamton University in New York have found a way to link images with the digital cameras that took them using the noise. The researchers applied for two patents this month and hope the technology will help prosecute child pornographers.
“The defense in these kind of cases would often be that the images were not taken by this person’s camera or that the images are not of real children,” said Jessica Fridrich, associate professor of electrical and computer engineering at Binghamton University, in the press release. “If it can be shown that the original images were taken by the person’s cell phone or camera, it becomes a much stronger case than if you just have a bunch of digital images that we all know are notoriously easy to manipulate.”
Fridrich and her fellow researchers, Jan Lukas and Miroslav Goljan, liken the technique to fingerprinting because each camera has its own unique noise pattern according to their research. In the study, the group’s technology looked at 2,700 images and linked them to 9 digital cameras with 100 percent accuracy. The technology has some limitations on it however. It requires either the camera that took the picture or multiple images for analyzing.
The technology is able to distinguish between manufacturers and models; the research group claims each digital camera has its own noise pattern “inevitably created during the manufacturing process,” according to the press release. The technology will be a hot topic in digital forgery cases. “We already know law enforcement wants to be able to use this,” Fridrich said. “What we have right now is a research tool; it’s a raw technology that we will continue to improve.”
